S-Plus is not alone in facing this issue. My wife had to
install SPSS because it was required by a class she's
taking. She needed to do a randomized blocks ANOVA.
By default, SPSS forces block*treatment interaction
in the model and she had to use a "model composer"
to override this. The net effect was using the
mouse in an incredibly tedious way to write out
a very simple additive effects formula.
Another example of GUIs causing problems: The
S-Plus 6.0 GUI for ANOVA does not automatically
convert factors to S factor objects, so if one
of your factors is an integer you will get
one degree of freedom for it no matter how many
levels it has. One purpose of a GUI should be to
prevent such user errors.
Another thing that makes me stay away from GUIs:
With web interfaces to systems, you can write
a web-"gui" to run the most commonly needed
analytical functions, whether running locally
or remotely. Such web interfaces use only
the computing/plotting components of the statistical
computing engine and bypass its GUI anyway.
Using JavaScript and other scripting languages
you can make multiple dynamic windows pop up
from the web browser as needed.
Frank Harrell
"Liaw, Andy" wrote:
>
> Perhaps what I should have said is: I gave up on the Splus GUI (both using
> and programming) long ago, and have no sympathy for those who struggle with
> it. (Sorry David.) IMHO (and I mean IMHO), it was a strategic mistake when
> MathSoft put the Axum GUI on Splus.
>
> Andy
>
> > -----Original Message-----
> > From: P.J.Wells@open.ac.uk [mailto:P.J.Wells@open.ac.uk]
> > Sent: Friday, September 21, 2001 10:29 AM
> > To: s-news@wubios.wustl.edu
> > Subject: Re: [S] Unix vs PC S Version problems
> >
> >
> > Andy Liaw wrote:
> >
> > >My guess is that you are using Splus functions written for the
> > >Splus GUI.
> > >If I'm not mistaken, these functions were intended to be
> > >called/used only
> > >through the Splus GUI, not from command line or scripts. In
> > so, you're
> > >hacking the system beyond its intended use, and I don't think
> > >Insightful can
> > >be blamed for not documenting those, as they may not be
> > >considered part of
> > >the API.
> >
> > Possibly, but my first contact with S-Plus was at an Insightful
> > demonstration where the ability to record actions performed
> > with the GUI and
> > then paste them into the command or script windows was
> > explicitly presented
> > as one of the program's selling points (it certainly sold it to me...)
> >
> > Julian Wells
> > OU Business School
> > The Open University
> > Walton Hall
> > Milton Keynes
> > MK7 6AA
> > United Kingdom
> > +44 1908 654658
> > ---------------------------------------------------------------------
> > This message was distributed by s-news@lists.biostat.wustl.edu. To
> > unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
> > the BODY of the message: unsubscribe s-news
> >
>
> ---------------------------------------------------------------------
> This message was distributed by s-news@lists.biostat.wustl.edu. To
> unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
> the BODY of the message: unsubscribe s-news
--
Frank E Harrell Jr Prof. of Biostatistics & Statistics
Div. of Biostatistics & Epidem. Dept. of Health Evaluation Sciences
U. Virginia School of Medicine http://hesweb1.med.virginia.edu/biostat
|